Introduction
Governor limits in Salesforce is one of the key considerations you need to follow while designingSOQL queries, DML statements, and etc in order to run applications without any issue. This article will help reduce CPU execution time–while processing records inside a “for” loop–to avoid reaching the CPU limit enforced by the Governor.
Processing Records in for Loop
Consider that you are going to access less than or equal to 50000 records and process these records in a loop. While processing these records in a “for” loop, the execution time will be longer depending upon your way of processing records. In order to achieve faster performance, you need to use a right approach in the “for” loop. Let us go through these four examples to understand how each approach takes different amount of execution time:
Example Code
public class LoopConsideration
{
public static void method1()
{
List<Account> accounts = [SELECT Id, Name FROM Account LIMIT 50000];
for(Integer i=0; i<accounts.Size(); i++)
{
accounts[i].Name = accounts[i].Name + 'Go';
}
}
public static void method2()
{
List<Account> accounts = [SELECT Id, Name FROM Account LIMIT 50000];
for(Account currAccount : accounts)
{
currAccount.Name = currAccount.Name + 'Go';
}
}
public static void method3()
{
for(Account currAccount : [SELECT Id, Name FROM Account LIMIT 50000])
{
currAccount.Name = currAccount.Name + 'Go';
}
}
public static void method4()
{
List<Account> accounts = [SELECT Id, Name FROM Account LIMIT 50000];
for(Integer i=0, j= accounts.Size(); i < j; i++)
{
accounts[i].Name = accounts[i].Name + 'Go';
}
}
}Run the above four methods using Developer Console to see the difference
Integer startTime = Limits.getCpuTime();
LoopConsideration.method1();
Integer endTime = Limits.getCpuTime();
System.debug('Spent Time'+ (endTime - startTime));Similarly, run remaining three methods by changing method name from LoopConsideration.method1() to LoopConsideration.method2(), LoopConsideration.method3(), and LoopConsideration.method4()
Here is the processing time for all four methods. For these examples, I have used 10000 records. Based on this record count, the process times are shown below:
Execution Time:
Method 1: Spent Time: 647 Method 2: Spent Time: 444 Method 3: Spent Time: 1558 Method 4: Spent Time: 368
Conclusion
The first three methods will check list size in each iteration and iterate components one by one, leading to more CPU consumption. So if you assign the list size to a variable, then processing records using that variable will take less time. Therefore; the method4() approach in the above class will reduce the CPU consumption.
Reference
